-
Notifications
You must be signed in to change notification settings - Fork 119
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Dorado 0.8.0 #1051
Add Dorado 0.8.0 #1051
Conversation
Nice work on this dockerfile 👍 OK the GitHub Actions runners don't have enough storage space to run the automated tests. That base image is pretty large which contributed to that. In lieu of the automated tests I've built the docker image locally to test it out. It builds successfully and the tests passes ✅ Requested changes:
One idea I had - users may not want to use a docker image that includes all models due to it's large size. They may be OK with just downloading the model at runtime and using that to save on docker image download time & storage space. They take up ~4.2GB:
What if we had a nearly identical dockerfile/docker image, but just remove the step to download all the model files? I think that might be useful to someone. Just a thought. We can do that in the future if someone requests it, not necessary to include with this PR. |
I have updated the relevant files with the required information and removed the lines not needed, thanks for the review! |
Thanks for making those requested changes, Fraser. I made a few small updates to clarify info in the dorado-specific readme.md file and updated some links to match formatting from other links. I was able to build the docker image locally, which does the cpu only test. It built successfully and tests passed. Here's the command I ran: I will test on a local computer with an NVIDIA GPU just to be sure this docker image will work on another machine with a GPU prior to approving/merging/deploying. |
OK I tested running this on a Windows 10 computer with WSL2 using ubuntu 20, with an NVIDIA 3080Ti GPU. Here's output of
Built the image locally with this command which succeeded:
Downloaded the test POD5 file with the
I was monitoring gpu activity with Double-checked the SAM file that was produced and it looks good:
Thanks for your patience with all the requested changes and testing. GPU-enabled software is not super common our collection of containers so I want to ensure it runs smoothly from the start 👍 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Approving, knowing that the automated tests failed due to resource limitations and that manual testing in external computing environments succeeded. Will deploy the docker image momentarily to dockerhub and quay using the image name staphb/dorado:latest
and staphb/dorado:0.8.0
as well as the versions with quay.io/
prefix
deploy workflow might fail, but we will see: https://github.com/StaPH-B/docker-builds/actions/runs/11109669669 If it does I can deploy manually to dockerhub and quay |
I deployed the docker image manually. The hard drive is filling up on the github actions runner. |
This pull request adds a Dockerfile and accompanying README for Dorado version 0.8.0.
Dorado is a high-performance, open-source tool developed by Oxford Nanopore Technologies for basecalling Oxford Nanopore sequencing data. It supports both CPU and GPU acceleration, providing rapid and accurate basecalling of Fast5/Pod5 files.
Key features of this Docker image:
Dorado Version 0.8.0: Includes the latest stable release of Dorado.
Pre-downloaded Basecalling Models: All necessary basecalling models are downloaded during the build process and stored in /dorado_models.
Sample Pod5 Test File: A sample Pod5 file is included for testing purposes.
Internal Test Stage: The Dockerfile includes an internal test stage that runs during the build process to ensure Dorado is installed correctly and functioning as expected.
NVIDIA CUDA Support: Based on the NVIDIA CUDA 12.2.0 base image to enable GPU acceleration (requires an NVIDIA GPU and the NVIDIA Container Toolkit).
The README.md for Dorado 0.8.0 is longer than 30 lines because it includes detailed instructions and explanations necessary for users to effectively build, run, and test the Docker image. Given that Dorado utilizes GPU acceleration, it's important to provide comprehensive guidance.
Demonstration of GPU Functionality
I have tested the Dorado 0.8.0 Docker image on a GPU-enabled Linux machine to confirm that it utilizes GPU acceleration as intended.
System Details:
GPU Model: NVIDIA Tesla T4
Driver Version: 460.32.03
CUDA Version: 12.2
Operating System: Ubuntu 20.04
nvidia-smi
Output:samtools view basecalled.sam
Pull Request (PR) checklist:
docker build --tag samtools:1.15test --target test docker-builds/samtools/1.15
)spades/3.12.0/Dockerfile
)shigatyper/2.0.1/test.sh
)spades/3.12.0/README.md
)